MASSACHUSETTS
INSTITUTE OF TECHNOLOGY
The RESEARCH LABORATORY of ELECTRONICS
Visual and Haptic Interactions in the
Perception of Stiffness in
Virtual Environments
By: Alberto J. Cividanes
and Mandayam A. Srinivasan
RLE Technical Report No. 640
June 2000
Visual and Haptic Interactions in the Perception of
Stiffness in Virtual Environments
By: Alberto J. Cividanes
and Mandayam A. Srinivasan
RLE Technical Report No. 640
June 2000
Visual and Haptic Interactions in the Perception of
Stiffness in Virtual Environments
by
Alberto J. Cividanes
Submitted to the Department of Mechanical Engineering
in partial fulfillment of the requirements for the degree of
Bachelor of Science
at the
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
June 2000
© 2000 Massachusetts Institute of Technology. All Rights Reserved.
Author ...................................................................................................
Department of Mechanical Engineering
May 5, 2000
Certified by ...........
...............................................................
Mandayam A. Srinivasan
Principal Research Scientist
Thesis Supervisor
Accepted by ...........................................................................................
Ernest E. Cravalho
Professor of Mechanical Engineering
Chairman, Undergraduate Thesis Committee
Visual and Haptic Interactions in the Perception
of Stiffness in Virtual Environments
by
Alberto J. Cividanes
Submitted to the Department of Mechanical Engineering on May 5,
2000, in partial fulfillment of the requirements for the degree of
Bachelor of Science in Mechanical Engineering
Abstract
Psychophysical experiments were conducted in order to determine the impact of manipulated visual information and 3-D perspective on the haptic perception of stiffness in virtual
environments. The Phantom force-reflecting haptic interface and a computer monitor were
utilized to run experiments on the human discrimination of stiffness of two virtual springs.
Ten subjects were asked to use the Phantom to compress two virtual springs sequentally
and feel the corresponding displacement and forces through their hand, while they
observed the visual deformation of the springs on a computer monitor. The subjects were
asked to judge which spring was softer. Without the knowledge of the subjects, the visually presented deformation of each spring was manipulated systematically across experimental trials so that its relationship to the haptic stiffness of the spring was varied.
Experiments were run on two different placement configurations of the springs. When the
springs were placed side by side, only the effect of this manipulated visual information
was studied. In a second configuration where the springs were placed one in front of the
other, the effect of 3-D perspective was added as another factor. Results show that for both
configurations, the percentage of correct responses decreased dramatically as the visual
scaling parameter increased, demonstrating a visual-haptic illusion in stiffness perception.
This suggests that visual information has a clear dominance over the kinesthetic hand
position information in the discrimination of stiffness in this type of virtual environments.
Thus, a proper combination of manipulated visual information and 3-D perspective can be
used to enhance the range of haptic experience in multimodal virtual environments. Both
the deliberately incorporated visual illusions and the ones created by 3-D perspective can
be exploited in order to compensate for the limitations that current haptic interfaces have.
Thesis Supervisor: Mandayam A. Srinivasan
Title: Principal Research Scientist
Acknowledgements
One of the most difficult tasks of writing this thesis was to enumerate all the people that
have helped with my research throughout the past year. First, I would like to thank my thesis and UROP advisor, Mandayam Srinivasan, who was given me his constant support and
advice for the last year and a half. Besides introducing me to the world of haptics, he has
been a great mentor who taught me the importance of conducting good scientific research.
I would also like to thank Wan-Chen Wu, a graduate student at the Touchlab. Without her
help with the C++ program, this study could never have been done. Also, I want to thank
her for her patience with me and for guiding me through all the minor details in this
research experience. I also have to thank almost everyone at the Touchlab for making this
such a nice experience. I cannot finish this section without mentioning those people who
contributed with the data by participating as subjects: Maria Isabel Carnasciali, Hugo Delgado, Luis Miranda, Fernando Perez, Jeriel Rivera, David Antolin Laguna, Dolores Cruz,
Suzie Mkandawire, Arin Basmajian, and Lizmarie Esparza. Finally, I wish to thank my
family who supported me and provided me with the most basic tools to succeed in life. I
hope you enjoy this thesis.
--_---··1·-11
1
__
1^__11_ 11
1_ ^11_11_._.-.-
__
--_1_11_11-·11
Table of Contents
1 Introduction . ............................................................................................................. 11
1.1 Purpose of the Study ........................................................
12
1.2 Haptic Interfaces and Virtual Environments .................................................... 12
2 Multisensory Research with Vision and Touch ....................................................... 15
3 Experiments
..
.....................................................
17
3.1 Apparatus ................ ..
......................................
17
3.2 Procedure ........................................................
18
3.2.1 Side - Side Configuration.......................................................................20
3.2.2 Rear - Front Configuration ........................................................
23
4 Results . .......................................................................................................... 25
5 Discussion ........................................................
33
6 Conclusions ........
......................... .......................
43
Appendix A Data Tables with Subject Responses ....................................................... 45
A.1 Subject Responses in Side - Side Configuration ..............................................45
A.2 Subject Responses in R - F Configuration: Front Spring Standard ................. 49
A.3 Subject Responses in R - F Configuration: Rear Spring Standard ................... 53
Appendix B Data Tables Apparent Stiffness of Springs ................................................ 57
B.1 Apparent Stiffness in S-S Configuration ....................................................... 57
B.2 Apparent Stiffness in R-F Configuration (F Standard) .................................... 59
B.3 Apparent Stiffness in R-F Configuration (R Standard) .................................... 61
B ibliography ...................................................................................................
63
5
I
-·
C------·---l
Illlsrll-
- -
·
·
I-----^---·-r-^--l
---
---
6
-------
I
_
III-·~
~ -iL~I-
__
·
·
^I
_
List of Figures
Figure 3.1: Phantom 1.5 ....................................
17
Figure 3.2: Experimental setup used in both experiments ............................................ 19
Figure 3.3: Side - Side configuration of virtual springs .....................................
Figure 3.4: Rear - Front configuration of virtual springs .
20
................................ 23
Figure 4.1: Percent correct versus X in S-S configuration .................................
26
Figure 4.2: Percent correct versus X in R-F configuration: Front spring standard ......... 27
Figure 4.3: Percent correct versus X in R-F configuration: Rear spring standard .......... 28
Figure 4.4: Percent correct versus X when AK/K
=25% .......................................
29
Figure 4.5: Percent correct versus X when AK/Ko=50% .......................................
30
Figure 4.6: Percent correct versus X when AK/Ko=75% ......................................
30
Figure 4.7: Percent correct versus X when AK/K o =100% .................................
31.......
Figure 5.1: Percent correct versus X when AK/Ko=25% .......................................
38
o
Figure 5.2: Apparent stiffness difference versus X when AK/K =25% .......................38
Figure 5.3: Percent of change in apparent stiffness difference when AK/K =25% ..... 38
Figure 5.4: Percent correct versus X when AK/K
o
=50%...................................
39
Figure 5.5: Apparent stiffness difference versus X when AK/K o =50% ....................... 39
Figure 5.6: Percent of change in apparent stiffness difference when AK/K o =50% ..... 39
Figure 5.7: Percent correct versus X when AK/K o =75% ............................................. 40
Figure 5.8: Apparent stiffness difference versus X when AK/K o =75% ....................... 40
Figure 5.9: Percent of change in apparent stiffness difference when AK/K
o
=75% ..... 40
Figure 5.10: Percent correct versus X when AK/K o =100% ......................................... 41
Figure 5.11: Percent correct versus X when AK/K o =100% ......................................... 41
7
-
-^-_ll·---Pillllll
--I^II.-·L--
~
1-_1
IY·-_I11-··
-
-
1-
-·---
--
Figure 5.12: Percent of change in apparent stiffness difference when AK/Ko=100%.41
8
_
__
_
L
_
___1_11
___
_
_ _I__
I
I
_I
List of Tables
Table A. 1: Percent correct versus X when AK/K =25% (S-S)......................................... 45
Table A.2: Percent correct versus X when AK/K =50% (S-S) ................................... 46
Table A.3: Percent correct versus X when AK/K =75% (S-S)......................................... 47
Table A.4: Percent correct versus X when AK/K o =100% (S-S) ....................................... 48
Table A.5: Percent correct versus X when AK/K, =25% (F Standard)............................. 49
Table A.6: Percent correct versus X when AK/K =50% (F Standard) ............................. 50
Table A.7: Percent correct versus X when AK/K =75% (F Standard) ............................. 51
Table A.8: Percent correct versus X when AK/K =100% (F Standard) ........................... 52
Table A.9: Percent correct versus X when AK/Ko =25% (R Standard) ............................. 53
Table A.10: Percent correct versus X when AK/K =50% (R Standard) ........................... 54
Table A. 11: Percent correct versus X when AK/K =75% (R Standard).......................... 55
Table A. 12: Percent correct versus X when AK/K =100% (R Standard) ......................... 56
Table B.1: Apparent stiffness of springs when AK/K =25% in S-S ................................. 57
Table B.2: Apparent stiffness of springs when AK/K =50% in S-S................................. 57
Table B.3: Apparent stiffness of springs when AK/K =75% in S-S ................................ 58
Table B.4: Apparent stiffness of springs when AK/K =100% in S-S ............................... 58
Table B.5: Apparent stiffness of springs when AK/K =25% in R-F (F Standard).......... 59
Table B.6: Apparent stiffness of springs when AK/K =50% in R-F (F Standard) ..........59
Table B.7: Apparent stiffness of springs when AK/K =75% in R-F (F Standard) .......... 60
Table B.8: Apparent stiffness of springs when AK/K =100% in R-F (F Standard) ........60
9
-_
1_~
_1
_
M
·
Table B.9: Apparent stiffness of springs when AK/K, =25% in R-F (R Standard) ..................... 61
Table B. 10: AApparent stiffness of springs when AK/K, =50% in R-F (R Standard) ................. 61
Table B. 1: Apparent stiffness of springs when AK/K, =75% in R-F (R Standard) .................... 62
Table B.12: Apparent stiffness of springs when AK/Ko =100% in R-F (R Standard) .................. 62
10
--_1111111111
--111_-.
Ill-1DIIYII^_I·LIUIXIII----L·---(-.
_._
111-__
_1~-C-
I
-
-
-
Chapter 1
Introduction
1.1 Purpose of the Study
The applications of haptic perception in virtual environments have increased dramatically during recent years. As virtual reality is increasingly used in the fields of entertainment, medicine, education and training, among others, a mere combination of visual and
auditory feedback is not enough to provide a full virtual reality experience. In order to take
full advantage of virtual reality, users must be able to touch, feel, and manipulate virtual
objects in addition to hearing and seeing them. Thus, haptic displays with force and tactile
feedback are necessary in order to improve the applications of virtual environments. It is
imperative to understand that in order to be successful, a virtual environment does not
need to perfectly replicate reality. Instead, it only needs to match the abilities and limitations of the human sensory, motor, and cognitive systems.
Intersensory research conducted throughout the years has revealed that visual information can alter the haptic perception of several spatial properties. As haptic interfaces are
not as advanced as those interfaces that merely provide visual and auditory cues, a proper
combination of visual and haptic cues can be used to enhance the impact of multimodal
virtual environments on a user.
This thesis is an attempt to study the effects that deliberately incorporated visual illusions and 3-D perspective can have on the perception of stiffness in these environments.
Psychophysical experiments have been conducted using the Phantom haptic interface, in
which subjects have been asked to discriminate between the stiffness of two virtual
springs. It is believed that a proper combination of manipulated visual information and 3-
11
_ II
_
IPII-.
^-·
----
-· II ii I
__
D perspective can be exploited in order to compensate for the limitations that current haptic interfaces have today in terms of bandwidth, resolution, and workspace.
1.2 Haptic Interfaces and Virtual Environments
At the pace at which computer technology is currently growing, visual and auditory
feedback alone cannot provide a person with a full experience in a virtual environment.
Haptic interactions with the virtual objects are essential in this type of environments, as
being able to touch, feel, and manipulate objects in virtual environments provides a sense
of immersion in the environment that is not possible with only visual and auditory cues. A
better immersion in a virtual environment can be achieved by the synchronous operation
of a visual and auditory display with a haptic interface.
Haptic interfaces enable manual interaction with virtual environments or teleoperated
remote systems. These devices are employed for tasks that are typically performed by
using the hands, primarily manual exploration and manipulation of objects [11]. These
manual interactions may also be accompanied by the stimulation of additional sensory
modalities such as vision and audition. As human users perform tasks with a haptic interface, desired motor actions are conveyed by physically manipulating the interface, which
displays tactual sensory information to the user by appropriately stimulating his or her tactile and kinesthetic sensory subsystems. Typically, force-reflecting haptic interfaces perform two basic functions: (1) To measure the positions, as well as their time derivatives, of
the users' hands and (2) to display contact forces to the user.
The haptic interface utilized in this experiment was the Phantom version 1.5. The use
of this device, developed by Sensable Technologies, has started a new field analogous to
12
_
1^~
1
_
1
___
_-
1~
1
---
I
II*---_
I
I
computer graphics called computer haptics. This new field is primarily related to the techniques and processes associated with generating and displaying synthesized haptic stimuli
to the human user [12].
The Phantom interface allows motion along six degrees of freedom and measures this
motion. In addition, it can exert controllable forces on the user along three of those
degrees of freedom. To evoke the sensation of touching objects, geometric, material, kinematic, and dynamic properties need to be modeled of the world that is desired to be represented. Rendering computer programs for the Phantom typically run at 500 to 2,000 Hz,
and it is important that these rendering algorithms are efficient.
13
.__I
-----------
^-------·sC----·LIL-u----
___-I1Yllllllllill·--·B·IC
1111_^-llln-^-
1-
_11~-·--·-·1
----
-11_-
-
-
--
·-
-
14
Chapter 2
Multisensory Research with Vision and Touch
Previous psychophysical research has been conducted studying the intersensory relationship between vision and touch. Evidence based on previous experiments shows that
visual information can alter the haptic perception of several spatial properties. Welch and
Warren have written an extensive review of experiments done in the past concerning
intersensory interactions, particularly visual and haptic interactions [16].
There are a few studies that are directly related to the one presented in this thesis. The
first one is a preliminary study conducted by Srinivasan, Beauregard, and Brock in 1996
that studied the impact of visual information on the haptic perception of stiffness in two
dimensional virtual environments [13]. By using a three degrees of freedom, force-reflecting haptic interface called the Planar Grasper, three subjects were asked to discriminate
between the stiffness of two virtual springs. The results were very consistent, and they
concluded that subjects tended to rely more on visual rather than haptic cues for judging
displacements. This behavior occurred even when the visual deformation of the springs
did not match their actual haptic deformation. Thus, they suggested the possibility that
manipulated visual information can be used to improve haptic perception experiences in
virtual environments.
In 1997, Durfee et al. conducted a similar experiment using a rotary electric motor as
the haptic interface [2]. Seven subjects were required to discriminate between the stiffness
of two virtual springs that had a mismatch between their visual and haptic stiffness. The
results show that when subjects made errors in haptic estimation, the errors tended to fol-
15
low the visual cues. This happened particularly for a large mismatch of visual and haptic
stiffness.
Another important study is closely related to the one presented in this thesis. An experiment conducted by Srinivasan, Wu, and Basdogan in 1999 studied the effect of 3-D perspective vision on the perception of size and object stiffness in virtual environments [18].
In this case, there was no deliberately introduced mismatch between visual and haptic
stiffness of the springs. This study is very important to the one on this thesis, as the same
virtual environment and experimental setup was used in both subjects, with the exception
that they did not manipulate the mismatch between visual and haptic stiffness. They concluded that the 3-D perspective visual displays generated visual and haptic illusions, such
that farther objects are perceived to be shorter in length when there are only visual cues
present, and softer when there are only haptic cues presence. However, when both sensory
cues were provided, sensory data was fused such that vision and touch compensated for
the bias due to each other.
16
11_1_1111_11__111_11l__l_----
-
Chapter 3
Experiments
3.1 Apparatus
A haptic device called the Phantom (Personal Haptic Interface Mechanism) version
1.5 was used in these experiments in order to render the haptic environment. The Phantom
is a three axis force reflecting device that provides six degrees of freedom; three active and
three passive (Figure 3.1). A stylus, which is a pen-like end effector, is held by the user
and manipulated to explore a 19cm x 27cm x 38cm haptic workspace. The maximum
exertable force with the Phantom is 8.5N. It has a maximum closed stiffness of 3.5 N/mm,
while the backdrive friction is 0.04N. The nominal position resolution for this haptic
device is about 0.03mm, and the inertia, or the apparent mass at the tip, is always less than
75g for any given configuration.
Figure 3.1: Phantom 1.5.
17
The graphic virtual environment was created using the OpenInventor computer graphics software. An SGI workstation (Indigo 2, 64 bit system and ISA interface card) was
connected to the haptic device through a 12-bit D/A and a 19in monitor was used to show
the visual environment. The monitor had a high resolution of 1280 x 1024. Both haptic
and graphic environments were synchronized and controlled by means of C programming.
The position and orientation of the tip of the stylus were sensed during the exploration
process, and a corresponding force signal was sent back to the device to construct the virtual objects specified by the computer program. The graphic display had an updating rate
of approximately 30 frames/s, while the haptic display had a rate of about I kHz. Both of
these rates exceeded the perceptual fusion frequencies for the human eye and hand,
respectively.
3.2 Procedure
Ten subjects between the ages of 20-22 years old participated in the experiments. They
were all undergraduate students at the Massachusetts Institute of Technology and were
paid on an hourly basis. Exactly five males and five females participated in the study, all of
which were right handed with no known hand disorders. All subjects used their right hand
for the experiments.
Two experiments were designed in order to investigate the effects of manipulated
visual information and 3-D perspective on the haptic perception of stiffness in virtual environments. In both cases, the subjects were asked to sit comfortably on a chair located at a
fixed distance from the computer and the haptic device (Figure 3.2). Two rectangular
blocks with slots on top to prevent the tip of the Phantom from slipping, were displayed to
the user and represented two virtual springs. By holding the stylus of the Phantom haptic
18
__I
^IIIIIX·ll
device, subjects were asked to press on the virtual blocks, as they saw on the monitor what
they were perceiving haptically with the Phantom. They were asked to discriminate
between the stiffness of the two virtual "springs" and answer to the question: "Which button is softer?". In the first experiment, the virtual blocks were placed side by side (S-S
configuration), and only the effect of manipulated visual information on the perception of
stiffness was studied. In the second experiment, the blocks were arranged in a Rear - Front
configuration (R-F configuration), where one was placed right in front of the other. In this
case, both the effects of visual information and 3-D perspective were studied. In both
experiments, a small black curtain was placed between the Phantom haptic interface and
the user so that the subjects could not readily observe the position of their hands.
/
i
-
...:
i;;
/S
I
1-
Figure 3.2: Experimental setup used in both experiments.
19
1--11-.11_·_1___-
-
XII^I __
3.2.1 Side - Side Configuration
In the S-S configuration, the two virtual springs shown as blocks were placed side by
side along the horizontal axis on the monitor screen and displayed to the subject. Figure
3.3 shows the actual screen that the subjects saw when the experiment was performed. The
subjects held the stylus of the Phantom and pressed the blocks within the slots at the top
while seeing what they touched on the monitor screen. A visual cursor showing the position of the stylus tip was displayed visually on the screen in order to help the subject navigate through the 3-D virtual environment.
Figure 3.3: Side - Side configuration of the virtual "springs".
As the subjects explored the virtual springs and felt the stiffness of each spring, they
were asked to observe the image on the screen that showed the corresponding spring
deformations. They were allowed to compress the springs as many times as they wanted,
and when they were finished they were asked to select which one of two springs felt softer
20
-~_X__III___
I1-I
--
---11·1--··---·1··_-·I
by answering to the question: "Which button is softer?". Subjects responded by typing "1"
for the left spring and "2" for the right spring. It should be noted that the subjects received
no performance feedback as they performed the experiment.
In each trial, the stiffness of one of the two springs was always equal to a reference
stiffness, Ko = 0.25N/mm. The stiffness of the other spring was always higher, the reference stiffness plus an increment (Ko+ AK). This AK increment was either 25%, 50%,
75%, or 100% of the reference stiffness. Trials were randomized so that both springs had
an equal probability of having the reference stiffness, and the subject could not have prior
knowledge of which spring was softer.
During the experiment, visual information was manipulated so that the visual deformation of the springs that the subject saw on the screen did not always correspond to the
actual or haptic deformation of the springs that they were sensing with their hand with the
aid of the Phantom. The discrepancy between the haptic deformation of the spring and the
visually displayed spring deformation was varied systematically, but was randomized
across the trials. It ranged from zero discrepancy, that is, that the visual deformation of
each spring corresponded to its haptic deformation, to a case of a completely interchanged
discrepancy, where the visual deformation of the softer spring was equal to the visual
deformation of the harder spring for that same force and vice versa. The relationship
between the haptic and visual deformation is given by the following set of equations:
K
Xh,ref
Xv, ref = (1 -
o
(3.1)
F
+F)K
(3.2)
F
Xh, comp
(3.3)
(Ko + K)
F
comp
(1 - X)(K, + AK) + XK
(3.4)
o
21
X_LI__III
_
I-
I _I
I
where Xh. ref and xv. ref are the haptic and visual displacements, respectively, for the
spring with the standard reference stiffness. The relations for
xh, rom
and xv,
omp
give the
haptic and visual displacements, respectively, for the comparison or stiffer spring (that is,
the spring with the reference stiffness plus the increment). The visual scaling parameter is
given by X.
It can be observed from equations 3.1 and 3.3 that actual spring deformations for a
given force are simply equal to the force divided by the stiffness of the spring. On the
other hand, equations 3.2 and 3.4 show that the spring displacements that are visually displayed on the monitor screen are equal to the applied force divided by a weighted average
of both spring constants. The influence of each spring stiffness constant depends of the
scaling factor, XA.
In the experiment, X was varied from 0 to 1 at intervals of 0.25. When
equals 0, the actual and visual deformations for each spring are the same. Increasing values of X skews visual information so that the softer spring is visually compressed less than
it should. For the case when
equals 0.5, both springs will thus have an identical visual
deformation, even though their stiffness are not equal. However, when
equals 1 the
visual displacement of each spring is determined only by the spring constant of the other
spring.
For this experiment, there were a total of 20 possible combinations of the four values
of increments in stiffness and the five values of X. Each experimental session consisted of
40 trials, that is, two trials for each possible combination. For each of the two trials per
case in every experimental session, the reference stiffness was contained by a different
spring. All the trials within a session were randomized, so no two experimental sessions
had the trials in the same particular order. For the S-S configuration, each subject was
tested on ten different experimental sessions, giving a total of 400 trials per subject (20 trials for each AK and X value pair).
22
_
II__I__L·____YII___^I
_CI
__-_1I---··-1II_---I
---Y·---rmDY-··--UIYY·······---·L-
(II-- PI-_IIC---·----IIY
3.2.2 Rear - Front Configuration
The experiment for the S-S configuration was repeated with a different configuration
of the virtual springs. In this case, the springs were positioned one in front of the other on
a rectangular block along the z-axis, so the spring in the front looked significantly larger
than the one in the back. Figure 3.4 shows the actual screen that the subjects saw when the
experiment was performed.
Figure 3.4: Rear - Front configuration of the virtual "springs".
In this case, the additional parameter of 3-D perspective was added to the experiment.
Just as in the experiment with the S-S configuration, the manipulated visual deformations
of the springs are still governed by equations 3.1 to 3.4.
For the R-F configuration, trials were arranged in a similar manner as in the S-S case.
However, since 3-D perspective is a new factor that needed to be taken into account, twice
23
-- IIIUII1_
_I*_^III___^_IIUY__I1_XILI_^I··LII·-I
__ IX-Crrm -
I
I
the amount of data sets were displayed to each subject. Thus, each subject had 400 trials
with the rear spring as the standard, and another 400 trials with the front spring as standard. The total of 800 trials per subject were arranged in 20 experimental sessions of 40
trials each. The S-S and R-F configuration experiments were alternated for every subject.
For every trial, the subject's response was recorded in order to compare it with the correct
answer to the question posed.
24
_
I_
IIPI^_IYI-II_
-II
I
ILCI-·I---·IX.
I-
--.-
---
--
I-
II--
1
I
Chapter 4
Results
The results from the experiments show that manipulated visual information has a great
impact on the perception of stiffness. Figures 4.1 to 4.3 show the results of both experiments graphically as plots of percent correct responses versus the visual scaling parameter
X. The results plotted are the averages across the ten subjects for the S-S configuration,
and for the R-F configuration with either the front or the rear spring as standard. The
results have also been tabulated numerically in Appendix A.
Figure 4.1 shows the results for the S-S configuration as percent of correct responses
as a function of X for all the values of AK. For the different cases of AK, the results are
very consistent when AK/K
o
is 50%, 75%, and 100%. However, for the case in which
AK/K o =25%, a more linear behavior can be observed.
The graph presented in figure 4.1 shows that for the S-S configuration, when
is zero
and the discrepancy between visual and haptic spatial information was nonexistent, subjects responded extremely well, getting between 90% to 98% of their responses correct.
These values approximate very closely the expected results. As X was increased, however,
the performance decreased for all subjects almost linearly as expected. At X=0.5, the percentage of correct responses ranged between 57% to 65%. This was the case when both
springs had exactly the same visual deformation for a given force. In the other extreme
case, when X was equal to 1, the subjects responded correctly only in 12% to 20% of the
trials.
25
___
_
_II__
ILtl___
_·_I_
_
__
·1
·IIll _
2
I - --
.
I
_-F-_
:
I~
-
....
.....-
: delta K/Ko=25%
- delta K/Ko=50%
0- delta K/Ko=75%
delta
___1
80
-
-
.
---a
lK/o=100%
-
:
.
70
60
t
8
50
1 1. .. . .
. . . ..
-.
.. .
............
. .. .. .. .
a-
40
.. ..
. .
. .
.. .. I . ..
....
30
20
. ....-
.
...
1.
---
...
..
,.*. .......
.
. .
.
..
.
.
A.......
%%.,.....
..
...
I........
10
t
n
C
::
. . . . . . . ..
.... .. .. ..
--- ----
I.......
J
0.1
I
I
0.2
0.3
I
I
0.4
0.5
lambda
Figure 4.1: Percent correct versus
I~~~~
I
AS
0.6
4
I
I
0.7
0.8
0.9
1
in S-S configuration
Similar results have been plotted for the experiments in the R-F configuration. Figures
4.2 and 4.3 show graphs in the form of percentage of correct responses versus the visual
scaling parameter X for the different cases of AK. Figure 4.2 shows the results for the case
when the front spring was the standard; that is, it contained the reference stiffness Ko . Figure 4.3 shows a similar graph for the R-F configuration, when the rear spring was the standard.
It can be observed from the graph presented in figure 4.2 that when the front spring
was the standard in the R-F configuration, the results show a similar behavior that in the SS configuration. When
=O, subjects responded correctly between 93% to 98% of the
--
cases. As X was increased, however, the performance decreased for all subjects as
expected. At X=0.5, the percentage of correct responses ranged between 63.5% when
26
-
~~~~~~~-- ~~~~~~~~~~~~~~~~~~~~~
I----------
. _ 0
.
AK/Ko=25% to 71% when AK/Ko=100%. At this point, both springs had exactly the
same visual deformation for a given force, and the difference in the visual perception was
due to the 3-D perspective. When X was equal to 1, the subjects responded correctly from
12% of the trials when AK/Ko=100% to 34.5% of the cases when AK/Ko =25%. Again, as
in the S-S configuration, the results were very similar for all different values of AK, except
when AK/K
o =25%.
101in
°
delta
delta
O- delta
* delta
--
910
..........
K/Ko=25%
K/Ko=50%
K/Ko=75%
K/Ko=100%
8To
. . . . . ...
:.. . . . . . . . . .... . . . . . ..
. .. . . . . .
70
0
0
6
- ---- - - - - - - -
0o
- -
.a)
00Q
5
0
'
'
'
4
. .. .. .. . .... . ... .
.... .. .. ..
.. .. .. .. ..... ... . .. . .. .. .. .. ..... .. .. .
I
3
2
.
. .
:' '. '
-- --- - ---:. - -- -- - ----
1
0
0.1
0.2
0.3
0.4
:
0.5
0.6
0.7
:
.. .... . .
0.8
s~~~~~~~~
0.9
1
lambda
Figure 4.2: Percent correct versus X in R-F configuration: Front spring standard
Figure 4.3 shows the results for the R-F configuration when the rear spring was the
standard. Again, results decreased steadily for all cases of AK/K o as X was increased.
However, for most values of X, the percentages of correct responses were consistently
lower than for the case in which the front spring was constant. Also, the behavior of the
curve for the case in which AK/K o =25% is similar to the corresponding curves in the S-S
27
-
---
·
·
·--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~..
_-------·
.
S-~------
-----
·---·--------
-
-
-----
·----·-----
·II111··
configuration and in the R-F configuration when the front spring was constant. Between
=0 and . =0.5 the number of correct responses was significantly lower than the other
-cases of AK/K,, but it was marginally higher when X was greater than 0.5.
When
-- 0, subjects responded correctly between 75% to 98% of the cases. For
A=0.25, correct responses varied from 65.5% to 77.5% across all cases. At X=0.5, the percentage of correct responses ranged between 31.5% and 49.5%, while it ranged between
15.5% and 29% when A=0.75. Finally, when X was equal to 1,the subjects responded correctly from 10% to 20% of the cases.
8
a
n
0
D.
I
lambda
Figure 4.3: Percent correct versus X in R-F configuration: Rear spring standard
28
_~~
I___~~~__
I
~
I
_
_·_1__1_1_
· 1_1__~~~~~~
11_
1_1
_11__1____11___1_^__·1_11_1--
I-II
-X---l-.
The same results have also been plotted separately for all the different cases of AK/K,.
These graphs are shown in figures 4.4 to 4.7. Each graph contains results for the S-S configuration, and both variations of the R-F configuration for their respective value of AK.
Figure 4.4 shows the results when the difference in stiffness was 25%. A similar graph is
shown in Figure 4.5 for the case when the difference in stiffness was 50%. Figures 4.6 and
4.7 show what happened when the stiffness differences were 75% and 100%, respectively.
All graphs contain error bars, representing the standard error of the mean.
10 U
I
3s~:a
.
9U,
I
:
-
I
x
Side-Side
F
standard
standard
.....................'-.......:-
.....................
. .... ...... .. ..... ..... ..... .. .. ..... ..... .....
i.
-- - F standard
810
-
0
a)
-5
7
...........
....
60 ).-··
..-
;O-
I -. I.................................
.:.................
r .............
.-. . ---
.
.
I............
. .......-
-
.........
-
.....
...
II)
4
3
2
1
:
:
:I
0.1
0.2
0.3
I\'. .
0
.
:.
'
.
0.4
.
0.5
lambda
.
.
.
.
.
.
.
.
.
.
.
0.6
Figure 4.4: Percent correct versus X when
0.9
0.7
0.8
AK/K o
=25%
1
29
---
·-
-
-
-
-
i
-rs
l~s-F
IIILI-·m-
...
,
_·-
·-P
__Y-lll·-
jsl^1·
_
"_
PF
'
..
_-L
^
II
P
I
-
-
90
'-
Side-Side
- - FR
-Rstmdvd
-*-
".',
-
stndrd
,'-
.N
a
70
60
I
0E 500
13..
40
-
...
.
30 -
-:
!-'.
.
..
.
of~~~~~~~~~~~~~~~~~~~~
"'
:
.i..
Xssts\>.
"\
:
.,
as~~~~~~~~~~~~~~~~~~~
. .X,_-:\
Il
'I
2n
0
-
I ·--
·· ·
-·i
·
I
0.
.
.
.
.
0.1
0.2
0.3
0.4
0.5
·
-
·
0.6
I
I
0.7
0.8
0.9
1
lambda
Figure 4.5: Percent correct versus x when AK/Ko=50%
i
14M
I
I-
l
k
90
.
: .
'
I . .
_.
...
. .. .. .. .. .. .. .
Sidkle-Side
.
...
.
-*-
F
.nd,,rd
-.e- R-stundwd
a~~~~a
80
70
N
.. . ...
.. .. .. .. .. . .. .. . . ... ..
.
.
.'.
'N
60
8
\
'E 50
a.
.
:
..
..
.
\
:
....
i
....
:
.. ,:
..........
'
..
-
40
\.
30
\·
..\...
..
\
·
\·
I
I.....·.
\T
·
·
=
·
20
10
.
..
_
0
C
I
J
0.1
&
0.2
...
.. .
I,
0.3
*
I
0.4
0.5
1
0.6
0.7
0.8
0.9
1
lambda
Figure 4.6: Percent correct versus x when AK/K,=75%
30
__
~ ~~ ~ ~
_
~~~~~C--~~~~
-
_IYIl__U___311__11_I_-·
-~
l_-~·
F-·-
---
I
I...
90 .......--.
..........
.90
.
-
.-
-|
I'-
.
Side-Side
r
u- I
2
-: - R standard
.
8 0 . ................
70O .....
.
. .. .. .-. ..
. .. .. ..
.. .. ... . ..
...
i\
60
60
.... .. .. . .. ...
30
.......
20.
.',
3o
l o - --- -- --- .. ... . .... .. .
0
00
'
0.2
0.1
"
"
.
... .. .. .
-... .. .
.
.... .. .' .. .... ... .. . ......
......
;
0.3
..... .
'
0.4
'
0.5
lambda
'
0.6
..........
'
0.7
........
0.8
0.9
1
Figure 4.7: Percent correct versus X when AK/K o =100%
These graphs provide another way of looking at the results, as it can be seen how the
number of correct responses varied for the different configurations under the same conditions. For all cases of AK/K,, percentages of correct responses tend to be higher for the RF configuration when the front spring was standard, followed by the results for the S-S
configuration. The results for the R-F configuration when the rear spring was standard
tended to be consistently lower than those for the S-S configuration across all values of
AK/K
.
31
-----11111111--Y_)----l
_..
w~-
·
U~I
CIIIC~·~ll·-l-I
----
I
-
_1·_·11
-·1
32
Chapter 5
Discussion
Psychophysical experiments were performed for both S-S and R-F configurations with
relative stiffness differences between the springs of 25%, 50%, 75%, and 100%. The
results of these experiments suggest that the manipulation of visual information can have a
strong impact in affecting the perception of mechanical stiffness in virtual environments.
The presence of visual dominance over the kinesthetic sense of hand position has become
evident, as the percentage of correct responses decreased dramatically as the visual parameter X increased for all cases of AK/K o in both configurations. The influence of 3-D perspective can also be described as significant, as for all cases of AK/K
o
the number of
correct responses when the front spring was the standard in the R-F configuration was
higher than the results for the S-S configuration as well as the case when the rear spring
was standard in the R-F configuration. Of the three configurations, the R-F configuration
with the rear spring as standard consistently provided the lowest number of percent correct
responses. Finally, it was also observed that the relationship between the percent of correct
responses and X was not symmetric about the 50% line of correct responses. As X=1
inverted the visual and haptic displacements of the two springs, in general it was expected
that for this value of X the percent of correct responses would tend to go to zero. However,
on average this value approximated 10% and not zero, while for X=0 the number of correct responses was close to 100% as expected.
The results obtained for both S-S and R-F configurations are very consistent, taking
into account that ten subjects participated in the study. In order to explain the results, a
concept we define as "apparent stiffness" must be introduced. The apparent stiffness of a
33
spring is defined as the actual applied force divided by the visual displacement of the
spring:
Kapparent,"
Kapparentn,r
x.
(5.1)
Xv=
(5.O2)
By substituting equation 3.2 into equation 5.1, and equation 3.4 into equation 5.2, the
following expressions are obtained for the apparent stiffness:
Kapparen,, = (1 -
X)K o + X(Ko + AK)
KapparentCo = (1 - )(K
(5.3)
(5.4)
+ AK) + XK o
where Kapparent,, is the apparent stiffness for the standard spring (the one with ::the true
is the apparent stiffness of the comparison spring.
reference stiffness Ko ) and Kapparent,,
Using these equations, the apparent stiffness was calculated for both springs in the S-S
configuration in order to obtain a possible explanation of what the subjects were perceiving under the influence of manipulated visual information. In the R-F configuration the
apparent stiffness was also calculated for both springs. However, equations:5.3 and 5.4
needed to be modified to account for the effect of 3-D. perspective. A proportionality constant of 1.33 was measured between the deformation of the front spring against that of the
rear one. As the rear spring seems visually smaller and thus seems to deform less than
what it actually should, the proportionality constant needed to be incorporated into equations 5.3 and 5.4. Thus, apparent stiffness for both springs in both cases of the R-F configuration, where the front and rear springs are standard, is given by the following set of
equations:
Kapparent,,,
KnpparentF ,
=
(1 -
= 1.33(1
)K, + X(K o + AK)
X-)(K o +AK) + 1.33.K
34
(5.5)
o
(5.6)
Kapparentre,
KapparentR com
where KapparentF
and
(5.7)
1.33(1 - )K o + 1.33X(K o + AK)
KapparentF,
= (1 - X)(K o + AK) + XKo
(5.8)
are the apparent stiffness for the standard and com-
parison springs when the front spring is the standard, and
KapparentRr4f
and
KapparentRo.p
the
respective apparent stiffness for the two springs when the rear spring is the standard.
Notice that equation 5.5 for
KapparentF-,.
and equation 5.8 for
Kapparenc..
remained the
same as equations 5.3 and 5.4, respectively.
The experiment with the S-S configuration only measured the effects that manipulated
visual information might have on the perception of stiffness, and did not include the
effects of 3-D perspective. In this case, if subjects had discriminated solely on the basis of
haptic force and displacement of their hand, and ignored all the provided visual information, then the subjects should have obtained nearly 100% correct, regardless of the value
of X.
However, the results for both S-S and R-F show that the subjects discriminated
between the stiffness of the two springs by taking into account both the visual and haptic
information provided. In order to predict what the behavior would be if the subjects had
responded based purely on the apparent stiffness, apparent stiffness values were calculated
for both springs for all cases of AK/K o . The difference between the apparent stiffness of
reference and companion springs was plotted as a function of X in order to find a relationship to the results presented in the graphs shown in figures 4.4 to 4.7. Similarly, the percent of change in apparent stiffness difference (difference in apparent stiffness divided by
the apparent stiffness of the reference spring) was plotted against X to look for similar
trends. All the values for apparent stiffness, apparent stiffness difference, and percent of
change in apparent stiffness difference have been tabulated in Appendix B. Values for
35
~~~~~~~~~~~__~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
~
_.CI
apparent stiffness difference, and percent of change in apparent stiffness difference were
plotted systematically across all values of AK/K,.
For each case of AK/K o, the obtained
values for the S-S configuration, R-F configuration with the front spring standard, and R-F
configuration with the rear spring standard were plotted in the same graph in order to
observe the differences between them.
Figure 5.1 shows the experimental results for the case when AK/Ko =25%. This graph
was already presented in figure 4.4, but is shown again so the similarities in the trends with
the graphs of apparent stiffness difference and percent of change in apparent stiffness difference versus
can be observed. Figure 5.2 shows the difference between the apparent
stiffness of the two springs as a function of
A
for the particular case when AK/K o =25%.
The graph on figure 5.3 shows the percent of change in apparent stiffness difference as a
function of X.
Figures 5.4 to 5.6 show the same set of graphs for the case when AK/K o =50%, while
figures 5.7 to 5.9 show the same results when AK/Ko=75% and figures 5.10 to 5.12 show
the same when AK/K,,=100%.
It was found that for the four values of AK/K o, there is indeed a strong relationship
between the experimental results and the predicted behavior in terms of apparent stiffness.
By observing the graphs showing the apparent stiffness difference versus A, it can be seen
that the percentage of correct responses was expected to decrease monotonically as the
visual scaling parameter X was increased. The effects of 3-D perspective are also captured
in these graphs, as the percent of correct responses was expected to be systematically
higher across values of
for the R-F configuration when the front spring was constant.
Also, the number of correct responses in the S-S configuration was expected to be higher
than in the R-F case when the rear spring was constant. This behavior can be observed in
the graphs showing the experimental results for all values of AK/K,,. However, the
36
expected behavior in terms of apparent stiffness differences is linear and skew-symmetric
about the horizontal axis. The experimental results do not show this complete linearity or
symmetry about the 50% correct horizontal axis. For all cases of
AK/K,
the experimental
curves seem to be flatter between X=0 and X=0.25 and approach 100% correct responses.
This can be justified, as 100% is the maximum percentage of correct responses that is possible. On the other hand, as X approaches 1, the results seem to approach 10% of correct
responses instead of the expected 0%. This behavior is very consistent for both configurations. However, the graphs of percent of change in apparent stiffness difference as a function of X do capture this lack of symmetry about the horizontal axis (Figures 5.3, 5.6, 5.9,
and 5.12). This suggests that the presence of lack of symmetry in the curves for experimental results is not a coincidence, but needs to be studied further.
Another observation made from the graphs is that as AK/K
o
increases, the slope of the
apparent stiffness difference curves increases. In the experimental results, it can be
observed how for the AK/Ko=25% the behavior seemed to be flatter. This observation is
best appreciated in figures 4.1 to 4.3. This behavior suggests how the impact of manipulated visual information is stronger for larger values of AK, and is another area that needs
further study.
37
I
I
Figure 5.1: Percent correct versus A when AK/K 0 =25%
I-e- F.idbd
04
K--~~~~~~~
01
e
I
-U
_-
----
2
-
0_
0
-02
_
03
04
05
08
07
08
09I
05
o0.
0.7
0.i
09
I
-02
I_
_
_,
-0.3
-04
2
002
01
03
0.4
Figure 5.2: Apparent stiffness difference versus
when AK/Ko=25%
-.- F is-gd.
20
Z
-
._
,
0
,
01
..
02
03
04
05
I.avvji
08
07
08
09
Figure 5.3: Percent of change in apparent stiffness difference when AK/K 0 =25%
38
__-
_
"e
100
80
01
0.2
0.3
04
0.5
06
07-
08
0-
0.1
05.4:
2
03
0.4
Iantda
0.5
0.6
0.7
0.8
0.9
70
0
1
Figure 5.4: Percent correct versus ?.when AK/ K =5 0 %
:
:
:
..
:~~~~~~~~
S'deSide
-- F
standard
R standard
.........................-
......................................
:
0.3
0.2
-. , .
2 O
C8 O
F
:
-
0.4
:
i~~~~~~~
.i -,
m :
:
-,_ _ ..
...
.
-
_--:
......................
.
.. . . .............................
.............................................
............
0
-o. 1
i
.............
....:. ...... .................. ....
-0.2
,:
:
: :
:
:
:
:
.,, ....-.-................................
.......................
...........
...........
-0.3
-0.4
_:
:
:
0.8
0.9
rnc
0
01
0.2
0.3
0.4
0.5
labrrda
0.6
0.7
t
Figure 5.5: Apparent stiffness difference versus X when AK/K o =50%
120
-:-
too
Fstanda.d
.
Rstandard
80
0 01
2e
00
02
03
04
...... ..........
'
:
:'
-i
.... .....
.......
E 40
3? 20
C2.O
I
-20
I
-40
o
0 1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
I
Figure
5.6:
stiffness
Percent
of
change
in
apparent
da
Figure 5.6: Percent of change in apparent stiffness difference when AK/K 0 =50%
39
--
----"--)-'
-
I
I
Figure 5.7: Percent correct versus X when AK/K, =75%
I
-02
3
-0.4
-0
001
0.2
05
0.3
0.4
0.5
0.e
.7
0.7
0.
0.9
1
1*rr
Figure 5.8: Apparent stiffness difference versus X when AK/Ko =75%
I
Figure 5.9: Percent of change in apparent stiffness difference when AK/K
o =75%
40
_
II
·
·
FC--~~~-
------PI--rrr--l-ulrCrYIII--·
-r·------
-* -_-_·--_--_-_____I___r_·____l-_r_-----·
lW
[I
I
90
I
-- .Sai
I
---F...d.
.
160
-
-
8 0
.
20
-;
..
... .....
ii; | | l t I i
10
O
02
0.1
03
04
05
0.6
0.7
0.8
0.9
1
lambda
Figure 5.10: Percent correct versus X when AK/K o =100%
*
0.5-
t·
·
s a
.
I
Ske-Ski.
. ...........
F
st ,I d..
..........................
0.4
0.23
.......
0.2
......
-
..-s
.......
-
,
------
........ ..... ..... ....
..........
.
I
4 . ... ..: ... .. .
..
-. _... . . ::.
::
:
0.7
0.8
El
t
6
-03 .......-
0.1
0
................
0.2
.. .......
0.3
0.4
.. .....-
0.5
a.mbda
.
..
0.6
0.9
1
Figure 5.11: Apparent stiffness difference versus X when AK/Ko=100%
lambda
Figure 5.12: Percent of change in apparent stiffness difference when AK/K o =100%
41
-
s-
111141(111 -----
-·-----·IC
YIIICYILIIII-I
_
F
b
42
II__
_
s
_
__
1*__1
111111_1_1_
~C-
I
y~~--
l
-- L1-il·
Chapter 6
Conclusions
Psychophysical experiments have been conducted on ten different subjects using the
Phantom haptic interface in order to determine if manipulated visual information and 3-D
perspective can influence the haptic perception of stiffness in virtual environments.
It was found that visual information has a clear dominance over the kinesthetic hand
position information in the discrimination of stiffness in this type of virtual environmnents. This dominance became more evident as the visual scaling parameter was
increased. In addition, 3-D visual perspective adds an important contribution to these illusions, making objects that are located farther away be perceived to be stiffer by the subjects. This can be concluded since for all cases of AK/K
o
the number of correct responses
when the front spring was the standard in the R-F configuration was higher than the S-S
configuration and the case when the rear spring was standard in the R-F configuration. Of
the three configurations, the R-F configuration with the rear spring as standard consistently provided the lowest number of correct responses, suggesting that the 3-D perspective makes objects that are farther away to be perceived as stiffer.
Finally, it was also observed that the relationship between the percent of correct
responses and X was not symmetric about the 50% line of correct responses. This lack of
symmetry was also observed in the graphs showing the percentage of apparent stiffness
difference versus X, and is an area in which further research needs to be done.
Further studies can also be designed in order to find up to what limits can manipulated
visual information be incorporated in visual environments in order to enhance the haptic
experience. It is believed that the visual scaling factor will reach a value where the illu-
43
UII
__
-.-·^-
_ _-l·llll--U-
11I-----_I.
· · 1_1
---*--_-_·---_____·__
1_1_1 ----I^
_ -_
sions will break down when the visual display may seem unrelated to the haptic experience.
In general, results from this study suggest that a proper combination of manipulated
visual information and illusions created by 3-D perspective can strongly influence the haptic perception of stiffness in virtual environments. Thus, by providing a proper relationship between haptic and visual information, and the addition of 3-D perspective, the haptic
experience in mutimodal virtual environments can be improved, overcoming the limitations of current haptic interfaces.
44
~~~~^
_
_··
_11^1
~~~~~~~_
1_1~~
_-
--
l~·_~--^-
__
1
4-
-
-_ -
-- --
-
Appendix A
Data Tables with Subject Responses
A.1 Subject Responses in Side - Side Configuration
X=0
X=0.25
X =0.5
X =0.75
X=1.0
Subject 1
90%
70%
45%
40%
15%
Subject 2
95%
75%
60%
15%
50%
Subject 3
95%
80%
70%
45%
5%
Subject 4
75%
80%
75%
35%
20%
Subject 5
90%
75%
40%
25%
15%
Subject 6
90%
80%
75%
65%
15%
Subject 7
80%
80%
45%
35%
20%
Subject 8
95%
90%
65%
60%
40%
Subject 9
100%
75%
45%
45%
5%
Subject 10
95%
80%
50%
0%
10%
90.5%
78.5%
57%
36.5%
19.5%
Average
Table A.1: Percent correct versus X when AK/Ko=25% (S-S)
45
----r I-
-----r-------
=0
=--0.25
X=0.5
=--0.75
X=1.0
Subject 1
95%
90%
45%
30%
0%
Subject 2
100%
100%
80%
30%
25%
Subject 3
100%
95%
60%
20%
5%
Subject 4
95%
95%
70%
50%
45%
Subject 5
100%
95%
55%
25%
0%
Subject 6
95%
85%
100%
25%
20%
Subject 7
95%
90%
55%
20%
0%
Subject 8
95%
95%
70%
50%
55%
Subject 9
100%
100%
50%
5%
5%
Subject 10
100%
90%
30%
0%
5%
Average
97.5%
93.5%
61.5%
25.5%
16%
Table A.2: Percent correct versus X when
AK/Ko=50%
(S-S)
46
~~ ~·IUI.·
·
·
Iyll~·
· ·
-pl
I·
_
-
----
X=0
X=0.25
X =0.5
A =0.75
X =1.0
Subject 1
100%
95%
65%
5%
0%
Subject 2
100%
100%
80%
30%
25%
Subject 3
100%
100%
75%
30%
5%
Subject 4
100%
85%
80%
75%
45%
Subject 5
100%
90%
55%
5%
10%
Subject 6
100%
95%
75%
25%
0%
Subject 7
100%
90%
55%
25%
0%
Subject 8
95%
95%
70%
40%
50%
Subject 9
100%
95%
50%
0%
0%
Subject 10
100%
100%
40%
0%
0%
Average
99.5%
94.5%
64.5%
23.5%
13.5%
Table A.3: Percent correct versus X when AK/K =75% (S-S)
47
A-0
A=0.25
X-0.5
X.=0.75
X=1.0
Subject 1
100%
100%
60%
15%
0%
Subject 2
100%
95%
70%
35%
10%
Subject 3
95%
75%
60%
15%
50%
Subject 4
95%
95%
80%
70%
20%
Subject 5
100%
95%
40%
5%
0%
Subject 6
100%
95%
60%
15%
10%
Subject 7
100%
95%
50%
15%
5%
Subject 8
100%
95%
65%
45%
50%
Subject 9
100%
100%
60%
15%
10%
Subject 10
100%
95%
25%
0%
0%
Average
99%
94%
57.5%
23%
15.5%
Table A.4: Percent correct versus
48
when AK/Ko =100% (S-S)
A.2 Subject Responses in R - F Configuration: Front Spring Standard
X=0
X=0.25
X=0.5
X=0.75
X=1.0
Subject 1
95%
95%
65%
45%
35%
Subject 2
90%
75%
60%
70%
70%
Subject 3
90%
90%
85%
80%
55%
Subject 4
90%
70%
65%
70%
25%
Subject 5
100%
70%
65%
60%
30%
Subject 6
100%
80%
65%
60%
35%
Subject 7
95%
80%
75%
45%
45%
Subject 8
80%
65%
70%
55%
30%
Subject 9
85%
75%
40%
20%
5%
Subject 10
100%
90%
45%
25%
15%
Average
92.5%
79%
63.5%
53%
34.5%
Table A.5: Percent correct versus X when AK/K o =25% (F Standard)
49
=0.75
--
X=1.0
X=0
X=0.25
X=0.5
Subject 1
100%
90%
80%
45%
15%
Subject 2
100%
95%
90%
60%
20%
Subject 3
95%
100%
75%
90%
35%
Subject 4
100%
95%
75%
65%
5%
Subject 5
100%
100%
75%
65%
70%
Subject 6
100%
100%
75%
45%
20%
Subject 7
100%
100%
50%
45%
5%
Subject 8
85%
95%
85%
65%
50%
Subject 9
100%
95%
40%
10%
0%
Subject 10
100%
95%
40%
10%
0%
Average
97.5 %
95%
70%
49.5 %
22.5%
Table A.6: Percent correct versus X when
50
AK/Ko
=50% (F Standard)
X=0
X=0.25
X--0.5
X-=0.75
X=1.0
Subject 1
100%
95%
65%
20%
10%
Subject 2
100%
100%
85%
60%
20%
Subject 3
95%
100%
95%
75%
40%
Subject 4
100%
90%
75%
65%
30%
Subject 5
100%
90%
80%
25%
5%
Subject 6
95%
100%
70%
45%
15%
Subject 7
100%
95%
60%
20%
0%
Subject 8
90%
60%
45%
45%
35%
Subject 9
100%
95%
60%
10%
0%
Subject 10
95%
80%
70%
10%
0%
97.5%
90.5%
70.5%
37.5%
15.5%
Average
Table A.7: Percent correct versus X when
51
AK/K o
=75% (F Standard)
X =0.75
X=1.0
65%
20%
5%
90%
85%
50%
25%
95%
100%
85%
55%
0%
Subject 4
95%
80%
95%
50%
45%
Subject 5
100%
100%
75%
20%
10%
Subject 6
90%
100%
80%
25%
5%
Subject 7
95%
95%
70%
20%
0%
Subject 8
95
85%
80%
45%
25%
Subject 9
100%
95%
30%
5%
0%
Subject 10
100%
90%
45%
10%
5%
Average
97%
93.5%
71%
30%
12%
A`=0
) =0.25
Subject 1
100%
100%
Subject 2
100%
Subject 3
--0.5
Table A.8: Percent correct versus A when
52
AK/Ko=100%
(F Standard)
A.3 Subject Responses in R - F Configuration: Rear Spring Standard
X=0
X=0.25
X =0.5
X =0.75
X=1.0
Subject 1
70%
65%
15%
20%
10%
Subject 2
90%
95%
50%
35%
50%
Subject 3
75%
50%
30%
30%
10%
Subject 4
45%
55%
35%
45%
40%
Subject 5
55%
45%
25%
10%
15%
Subject 6
75%
65%
30%
40%
25%
Subject 7
95%
50%
45%
20%
5%
Subject 8
60%
55%
25%
45%
35%
Subject 9
90%
90%
30%
30%
5%
Subject 10
95%
85%
30%
15%
0%
Average
75%
65.5%
31.5%
29%
19.5%
Table A.9: Percent correct versus X when AK/K o =25% (R Standard)
53
=0
=0.25
---
--0.5
=--0.75
X=1.0
Subject 1
90%
45%
35%
30%
15%
Subject 2
95%
85%
60%
25%
50%
Subject 3
85%
70%
35%
10%
0%
Subject 4
80%
50%
60%
25%
30%
Subject 5
85%
75%
5%
15%
5%
Subject 6
95%
70%
40%
20%
10%
Subject 7
100%
65%
35%
10%
5%
Subject 8
90%
60%
45%
45%
35%
Subject 9
100%
80%
55%
10%
0%
Subject 10
100%
90%
30%
10%
0%
Average
92%
69%
40%
20%
15%
Table A.10: Percent correct versus
54
when
AK/Ko=50% (R
Standard)
X=0
X=0.25
X =0.5
X=0.75
X=1.0
Subject 1
85%
80%
20%
5%
10%
Subject 2
95%
90%
70%
45%
25%
Subject 3
90%
40%
25%
10%
0%
Subject 4
90%
80%
65%
15%
20%
Subject 5
100%
55%
20%
5%
0%
Subject 6
100%
70%
50%
5%
0%
Subject 7
100%
90%
60%
15%
5%
Subject 8
80%
80%
60%
40%
30%
Subject 9
100%
90%
55%
5%
5%
Subject 10
95%
80%
70%
10%
0%
93.5%
75.5%
49.5%
15.5%
9.5%
Average
Table A.11: Percent correct versus X when AK/K o =75% (R Standard)
55
A=0
,A=0.25
A =0.5
A =0.75
A=1.0
Subject 1
100%
65%
20%
15%
0%
Subject 2
95%
85%
60%
35%
30%
Subject 3
95%
85%
25%
5%
0%
Subject 4
95%
65%
60%
30%
20%
Subject 5
100%
60%
10%
5%
5%
Subject 6
100%
75%
25%
10%
0%
Subject 7
100%
90%
50%
10%
5%
Subject 8
95%
70%
60%
45%
20%
Subject 9
100%
85%
30%
15%
5%
Subject 10
100%
95%
30%
0%
5%
Average
98%
77.5%
37%
17%
9%
Table A.12: Percent correct versus A when
AK/Ko
=100% (R Standard)
56
--11-----~1^-~'-'
' '' ----···1111111
Appendix B
Data Tables with Apparent Stiffness of Springs
B.1 Apparent Stiffness in S-S Configuration
Apparent
Stiffness
(Standard)
[N/mm]
Apparent
Stiffness
(Comparison)
[N/mm]
Apparent
Stiffness
Difference
[N/mm]
0
0.25
0.3125
0.06
25%
0.25
0.265625
0.296875
0.03
12%
0.5
0.28125
0.28125
0
0%
0.75
0.296875
0.265625
-0.03
-11%
1.0
0.3125
0.25
-0.06
-20%
Difference
Table B.1: Apparent stiffness of springs when AK/Ko=25% in S-S
Apparent
Stiffness
(Standard)
[N/mm]
Apparent
Stiffness
(Comparison)
[N/mm]
Apparent
Stiffness
Difference
[N/mm]
Difference
0
0.25
0.375
0.13
50%
0.25
0.28125
0.34375
0.06
22%
0.5
0.3125
0.3125
0
0%
0.75
0.34375
0.28125
-0.06
-18%
1.0
0.375
0.25
-0.13
-33%
~.
Percent
of
Table B.2: Apparent stiffness of springs when AK/K o =50% in S-S
57
Apparent
Stiffness
(Standard)
[N/mm]
Apparent
Stiffness
(Comparison)
[N/mm]
Apparent
Stiffness
Difference
[N/mm]
Percent
0
0.25
0.4375
0.19
75%
0.25
0.296875
0.390625
0.09
32%
0.5
0.34375
0.34375
0
0%
0.75
0.390625
0.296875
-0.09
-24%
1.0
0.4375
0.25
-0.19
-43%
Difference
Table B.3: Apparent stiffness of springs when AK/Ko =75% in S-S
Apparent
Stiffness
(Standard)
[N/mm]
Apparent
Stiffness
(Comparison)
[N/mm]
Apparent
Stiffness
Difference
[N/mm]
Percent
0
0.25
0.5
0.25
100%
0.25
0.3125
0.4375
0.13
40%
0.5
0.375
0.375
0
0%
0.75
0.4375
0.3125
-0.13
-29%
1.0
0.5
0.25
-0.25
-50%
of
erence
Table B.4: Apparent stiffness of springs when AK/K,,=100% in S-S
58
-
.
. ,~~~~~1~-
11~
B.2 Apparent Stiffness in R-F Configuration (F Standard)
Apparent
Stiffness
(Standard)
[N/mm]
Apparent
Stiffness
(Comparison)
[N/mm]
Apparent
Stiffness
Difference
[N/mm]
Percent
0
0.25
0.416667
0.17
67%
0.25
0.265625
0.395833
0.13
49%
0.5
0.28125
0.375
0.09
33%
0.75
0.296875
0.3541667
0.06
19%
1.0
0.3125
0.3333
0.02
7%
Difference
Table B.5: Apparent stiffness of springs when AK/K, =25% in R-F (F Standard)
X
Apparent
Stiffness
(Standard)
[N/mm]
Apparent
Stiffness
(Comparison)
[N/mm]
Apparent
Stiffness
Difference
[N/mm]
Percent
0
0.25
0.5
100%
0.25
0.28125
0.458333
63%
0.5
0.3125
0.416667
33%
0.75
0.34375
0.375
9%
1.0
0.375
0.3333
-11%
of
Difference
Table B.6: Apparent stiffness of springs when AK/K o =50% in R-F (F Standard)
59
-"-
d"
- I_.__CI*UIIIIII--P--··
XICI1-IIII---l---I_II^F
11
II--
Apparent
Stiffness
(Standard)
(N/mm]
Apparent
Stiffness
(Comparison)
Apparent
Stiffness
Difference
Percent
of
[N/mm]
[N/mm]
0
0.25
0.58333
0.33
133%
0.25
0.296875
0.520833
0.22
75%
0.5
0.34375
0.458333
0.11
33%
0.75
0.390625
0.395833
0.01
1%
1.0
0.4375
0.3333
-0.10
-24%
Table B.7: Apparent stiffness of springs when AK/K, =75% in R-F (F Standard)
Apparent
Stiffness
(Standard)
[N/mm]
Apparent
Stiffness
(Comparison)
[N/mm]
Apparent
Stiffness
Difference
[N/mm]
Percent
0
0.25
0.6667
0.42
167%
0.25
0.3125
0.58333
0.27
87%
0.5
0.375
0.5
0.125
33%
0.75
0.4375
0.41667
-0.02
-5%
1.0
0.5
0.333
-0.17
-33%
Difference
Table B.8: Apparent stiffness of springs when AK/Ko=100% in R-F (F Standard)
60
~~~~~~~~~~~~~~~~_
..
-I-
_
.
|
.
.
.
.~~
_
^
B.3 Apparent Stiffness in R-F Configuration (R Standard)
X
Apparent
Stiffness
(Standard)
[N/mm]
Apparent
Stiffness
(Comparison)
[N/mm]
Apparent
Stiffness
Difference
[N/mm]
of
0
0.3333
0.3125
-0.02
-6%
0.25
0.3541667
0.296875
-0.06
-16%
0.5
0.375
0.28125
-0.09
-25%
0.75
0.395833
0.265625
-0.13
-33%
1.0
0.41667
0.25
-0.17
-40%
Table B.9: Apparent stiffness of springs when AK/K o =25% in R-F (R Standard)
Apparent
Stiffness
(Standard)
[N/mm]
Apparent
Stiffness
(Comparison)
[N/mm]
Apparent
Stiffness
Difference
[N/mm]
Percent
0
0.3333
0.375
0.04
13%
0.25
0.375
0.34375
-0.03
-8%
0.5
0.41667
0.3125
-0.10
-25%
0.75
0.45833
0.28125
-0.18
-39%
1.0
0.5
0.25
-0.25
-50%
Table B.10: Apparent stiffness of springs when AK/K o =50% in R-F (R Standard)
61
--
..-
*
._
u
_
-
· *11
_
-
_
___s
·^
Apparent
Stiffness
(Standard)
[N/mm]
Apparent
Stiffness
(Comparison)
[N/mm]
Apparent
Stiffness
Difference
[N/mm]
Percent
0
0.3333
0.4375
0.10
31%
0.25
0.395933
0.390625
-0.01
-1%
0.5
0.45833
0.34375
-0.11
-25%
0.75
0.520833
0.296875
-0.22
-43%
1.0
0.58333
0.25
-0.33
-57%
Difference
Table B.11: Apparent stiffness of springs when AK/K, =75% in R-F (R Standard)
.
Apparent
Stiffness
(Standard)
[N/mm]
Apparent
Stiffness
(Comparison)
[N/mm]
Apparent
Stiffness
Difference
[N/mm]
Percent
of
Difference
0
0.3333
0.5
0.17
50%
0.25
0.41667
0.4375
0.02
5%
0.5
0.5
0.375
-0.13
-25%
0.75
0.58333
0.3125
-0.27
-46%
1.0
0.66667
0.25
-0.42
-63%
Table B.12: Apparent stiffness of springs when AK/Ko=100% in R-F (R Standard)
62
References
[1] Burdea, G.C. 1996, Force and Touch Feedback for Virtual Reality, John Wiley and
Sons, Inc.
[2] Durfee, W. K., Hendrix, C.M., Cheng, P., Varughese, G., 1997, "Influence of Haptic
and Visual Displays on the Estimation of Virtual Environment Stiffness", ASME
Dynamic Systems and Control Division, DSC-Vol. 61, pp. 139-144, Fairfield, New
Jersey.
[3] Gliner, C. R., Pick, A.D., Pick, H. L., Hales, J.J., September 1969, "A Developmental
Investigation of Visual and Haptic Preferences for Shape and Texture", Monographs of
the Societyfor Research in Child Development, Serial No. 130, Vol. 34, No. 6.
[4] Heller, M. A., 1982, "Visual and Tactual Texture Perception: Intersensory Cooperation", Perception & Psychophysics, Vol. 31, pp. 339-344.
[5] Heller, M. A., Schiff, W., 1991, The Psychology of Touch, Lawrence Erlbaum Associates, Inc.
[6] Kinney, J. A., Luria, S. M., 1970, "Conflicting Visual and Tactual-Kinesthetic Stimulation", Perception & Psychophysics, Vol. 8, pp. 189-192.
[7] Lederman, S. J,. Taylor, M. M., 1969, "Perception of Interpolated Position and Orientation by Vision and Active Touch", Perception & Psychophysics, Vol. 6, pp. 153-159.
[8] McDonald, L., Vince, J., 1994, Interacting with Virtual Environments, John Wiley
and Sons, Inc.
[9] Rosenberg, L. B., Adelstein, B. D., 1993, "Perceptual Decomposition of Virtual Haptic Surfaces", Proceedingsof IEEE Symposium on Research Frontiersin Virtual Reality, pp. 46-53, San Jose, California.
[10] Salisbury, J. K., Srinivasan, M. A., Sept. - Oct. 1997, "PHANToM Based Haptic
Interaction with Virtual Objects", IEEE Computer Graphicsand Applications.
[11] Srinivasan, M. A., 1995, "Haptic Interfaces", in Durlach, N. I., and Mavor, A. S.
(Eds.), Report of the Committee on Virtual Reality Research and Development, Chap.
4, National Research Council, National Academy Press, Washington, D.C.
[12] Srinivasan, M. A., Basdogan, Cagatay, 1997, "Haptics in Virtual Environments: Taxonomy, Research Status, and Challenges", Computers and Graphics, Vol. 21, No. 4.
[13] Srinivasan, M. A., Beauregard, G.L., Brock, D.L., 1996, "The Impact of Visual Information on Haptic Perception of Stiffness in Virtual Environments", Proceedings of
ASME Dynamic Systems and Control Division, DSC-Vol. 58, pp. 555-559, Atlanta,
Georgia.
[14] Stuart, R., 1996, The Design of Virtual Environments, McGraw-Hill.
[15] Tan, H. Z., 1997, "The Identification of Sphere Size Using the PHANToM: Towards
a Set of Building Blocks for Rendering Haptic Environment", Proceedings of ASME
Winter Annual Meeting.
[16] Warren, D.H., and Welch, R.B., 1986, "Intersensory Interactions", Handbook of Perception and Human Performance, Vol. I, Sensory Processes and Perception, Chapter
25, Wiley and Sons.
63
_____
[171 Weisenberger, J. M., Krier, M. J., 1997, "Haptic Perception of Simulated Surface
Textures via Vibratory and Force Feedback Displays", ASME Dynamic Systems and
Control Division, DSC-Vol. 61, pp. 55-60, Fairfield, New Jersey.
[18] Wu, W. C., Srinivasan, M.A., and Basdogan, C., 1999, "Visual, Haptic, and Bimodal
Perception of Size and Stiffness in Virtual Environments", ASME Dynamic Systems
and Control Division, DSC-Vol. 67.
64
--------
_-~-1~·--11 ~II-·-·/Pli~
~1 ~ ____
._-
-
I
-_
^~~~~~~-
65
___
___II
IIII-_II--ILPI
1III-
I_
66
-
-
-
·i
_I_
__
__
67
__111·
11··
IUI----_II---I-_L-
1 1..--^·-·1·--1--1·----·-11_1_-1
ltllllYLX
+
-a
68
11_
_lllI
_II-
_
~
-_
__
I
_
_
69
_____1__1
_
_I___
IPILI·II·X^II1·YI---
--111111--·1__
-__
_Il--_
-.
1_
-_~_
--
1~